1. Field of the Invention
The present invention relates to pulsed electromagnetic sensors, and more particularly to a pulse center detector (PCD) for pulse echo radar and time-domain reflectometry (TDR) sensors. These sensors can be used for rangefinding, for automation, or for determining the fill-level of a tank.
2. Description of Related Art
Radar echo pulses exhibit large amplitude variations, depending on target size, range and dielectric constant, and these variations produce range measurement errors when the pulses are detected with a fixed threshold detector. Echo amplitude variations also occur to a lesser extent with TDR-based tank level sensors, mainly being limited to dielectric constant variations of the liquid in the tank. However, accurate TDR-based tank level sensors require accurate, amplitude-independent pulse detectors.
Detectors with amplitude-tracking thresholds or other means to achieve amplitude independence are generally termed automatic pulse detectors and several automatic pulse detectors have been in existence for more than 30 years. U.S. Pat No. 5,610,611, High Accuracy Material Level Sensor, to McEwan, describes the well-known constant fraction discriminator, or CFD, for use in a TDR-based tank level sensor. The CFD in the ""611 patent uses a peak detector to determine the peak amplitude of repetitive equivalent-time pulses and sets a trigger point that is a fraction of the peak amplitude, such as the half-way point on the rise of the pulse (half-max detection). Unfortunately, the CFD exhibits latency errors caused by slow peak tracking when the pulses decrease in amplitude. Latency is a particular problem when the CFD is used in a TDR level sensor for sloshing liquids in a tank, such as an automotive gas tank. Another potential problem with the peak detector is that it will lock-on to the strongest peak in a radar or TDR waveform, such as the main bang peak, and thus create errors in echo pulse detectionxe2x80x94unless the CFD is provided with analog gating, or separate, gated CFD""s are used for the transmit and echo pulses. Yet another problem with the CFD is that it will trigger on baseline noise when no echo pulses are present, so a threshold detector is needed to inhibit operation on weak pulses. To fully overcome all the limitations of a CFD, substantial additional circuitry is needed.
Another well-known automatic pulse detector is the time-of-peak (TOP) detector. The TOP detector differentiates pulses and triggers on the resulting zero axis crossings. To prevent false triggering on baseline noise, the desired pulses must be above a threshold before zero axis detection is enabled. This standard detector is utilized in an application to TDR in U.S. Pat. No. 5,457,990, Method and Apparatus for Determining a Fluid Level in the Vicinity of a Transmission Line, to Oswald, 1995. However, the TOP detector is less accurate than the CFD for the simple reason that a pulse peak is somewhat flat and has a low rate of change, making accurate time-of-peak detection difficult. Small baseline perturbations, such as baseline ringing or radar clutter, will sum with the pulse and substantially displace the exact time-of-peak. In contrast, a CFD detects at a fast slewing point of the pulse risetime, so the detection time is much less sensitive to baseline perturbations.
An amplitude-independent pulse detector is needed that (1) triggers on a high-slew point of a pulse like the CFD to avoid the inaccuracies of the TOP detector, (2) does not have the latency of the CFD, and (3) does not have the complexity of the prior automatic detectors.
According to the invention, a pulse center detector (PCD) threshold detects radar or TDR baseband transmit and echo pulses using a single fixed threshold comparator to produce transmit and echo detection pulses. The leading edges of the detection pulses are then formed into a leading-edge PWM (pulse width modulation) pulse (or xe2x80x9cleading-PWM pulsexe2x80x9d) having a width proportional to the time between the leading-edge of the transmit pulse and the leading-edge of the echo pulse, i.e., the transmit-echo range. Similarly, a trailing-edge PWM pulse (or xe2x80x9ctrailing-PWM pulsexe2x80x9d) is formed from the trailing edges of the detection pulses.
The leading- and trailing-PWM pulses gate separate range counters. The range counts are added and divided by two to produce an averagexe2x80x94or midpointxe2x80x94count that corresponds to a PWM pulse centered on the baseband transmit and echo pulse centers. Thus, pulse center detection is achieved, i.e., the range is measured from the center of the transmit pulse to the center of the echo pulse.
By definition, a pulse center is the midpoint between its leading and trailing edge. The center of a pulse does not necessarily correspond to the peak of the pulse, but it often does. The PCD relates the pulse center to the time average of the leading and trailing edge times of the pulse.
An analog alternative to digital count summation involves simple resistive summing and integration of the leading-PWM and trailing-PWM pulses to produce a DC value corresponding to the time difference between the centers of the baseband transmit and echo pulses. This DC value is a range-proportional analog output voltage. Analog summation and integration is a simple approach suitable for driving analog xe2x80x9cgas gauges.xe2x80x9d With either analog or digital summation, the final range indication is related to the centers of the transmit and echo pulses and does not vary with pulse amplitude. Although baseband pulses generally exhibit symmetric rise and fall times, asymmetric rise and fall times can be compensated by weighted PWM addition.
Timing jitter is lower in the PCD compared to prior automatic detectors since the addition of the leading and trailing PWM pulses averages the noise and yields a 3 dB reduction in noise, or equivalently, timing jitter.
The PCD can enhance the accuracy and reduce the cost of TDR-based electronic dipsticks for automotive gas gauges, industrial vat level sensors, automatic swimming pool regulators, and toilet tank level controllers. It can also be used to improve the performance of radar rangefinders for robotics or for automotive backup warning.